A new technique for determining sea surface temperatures (SST), based on the unsaturation ratio, U37k, of long chain C37 methyl alkenones produced by the phytoplankton Prymnesiophyceae, has previously been shown by other researchers to have a linear response with temperature in culture and water column studies, and a good correlation with delta-O-18 over approximately 100,000 years in a core from the east Equatorial Atlantic. These results indicate this technique has potential for determining paleo-SST in marine sediments. In order to apply U37k rigorously and quantitatively in sediments for paleotemperature estimations, we have calibrated the method in sediments and evaluated how well the alkenone's temeprature signal is preserved under some common conditions of deposition and sample handling, i.e. the effects on these compounds of dissolution of sedimentary carbonates, and storage and sampling. Comparison of U37k in duplicate samples collected and stored frozen versus those stored at room temperature for up to four years shows no significant differences. Carbonate dissolution which renders foraminifera-based techniques unusable or unreliable show no effect on U37k values in laboratory experiments. Initial field results support this, but more studies are necessary. The U37k "thermometer" was calibrated by analyzing U37k in core-tops from widely varying open ocean sites. Sediment values of U37k reflected overlying SST, with the results falling on the same regression line for culture and water column samples of Prahl and Wakeham [1], indicating that their equation (U37k = 0.033 T + 0.043) is suitable for use in converting U37k values in sediments to overlying warm season SST. Neither a wide range of productivity nor lowered salinity of the overlying waters appears to affect U37k accuracy. With a calibration for sediments, the U37k method can begin to be quantitatively applied down core to open ocean sediments for paleotemperature estimations of the late Quaternary.